The manner in which data was collected and analyzed heavily influences it’s (evidential) value. When decisions regarding analyses and publication are conditional on the data (e.g., reporting or deciding to publish a result because it fits with a theory) this distorts the scientific literature. A powerful way to circumvent this source of bias is making such decisions before the results are known. There are the moment two ways to do this:
- Independently pre-register the methodology of a study before you collect any data; you can show the time-stamped pre-registration when you submit your study to a journal.
- Registered Report: Pre-registering a study and having it peer-reviewed and accepted for publication at a journal. In addition to the massive advantage that this completely removes publication bias, this has the advantage of reviewers being able to potentially improve the quality of the method before the study has taken place.
Useful and information websites:
AsPredicted.org – A very easy to use website where you can pre-register your studies.
Open Science Framework – Pre-register your studies, and share your pre-prints, data, and materials
Registered Reports – Contains a good explanation of pre-registration, a list of all journals offering Registered Reports, templates for editors and authors, and more.
List of all Registered Reports & pre-registered studies – A Zotero library which attempts to keep track of all pre-registered studies
List of all Registered Reports & pre-regsitered studies in education – A subset of the above library which lists only studies in education.